The present invention relates to a method and an apparatus for fabricating an evaporation device having semiconductor characteristics in an active (or atomic state) gas atmosphere.
In case that an evaporation device of this kind, for example, a photoconductive target for a television camera tube, which is formed by a hetero-junction of material such as CdS or CdTe consisting of compounds having elements in IIb and VIb groups of the periodic table, is fabricated by an ordinary high vacuum deposition method, the surface of a film thus fabricated has better uniformity and graininess and higher mass productivity than a target film fabricated by other well-known deposition methods, for instance, a vapor phase reaction method or sputtering method. However, such characteristics as dark current, resolution, residual image and image burning are remarkably inferior.
The reasons for the above will first be described from a viewpoint of crystallography.
(1) A target film fabricated by an ordinary high vacuum deposition method has a structure in the form that fine crystals are deposited on a substrate. Individual crystal has a size, namely, a grain size (G.S.) of less than 300 A at most. Accordingly, many grain boundaries (G.B.) lie between those fine crystals forming the target film, so that the transfer of signal charges are prevented. This results in the residual image and the image burning frequently.
(2) Segregated component atoms which do not grow up yet to a semiconductor compound, for instance, Te atoms in CdTe or Cd atoms in CdS, which will be referred to as deposits, hereinafter, exist on the grain boundaries. These deposits consist of metal or semi-metal. Accordingly, the inner parts of the deposited film are shortcircuited by these deposits, so that an electric resistance is extremely reduced in a direction perpendicular to the film surface, namely, in the thickness direction and in a direction parallel to the film surface, namely, in the surface direction. As a result, the resolution and dark current characteristics are remarkably deteriorated.
(3) The above-mentioned deposits prevent remarkably the growth of the fine crystals. As a result, the above-mentioned grain size is reduced, and the above-mentioned grain boundaries are increased.
In order to remove those disadvantages of the high vacuum deposition method, various compensating procedures have been attempted as follows.
(a) The substrate temperature is increased during evaporation.
(b) After the evaporation, the recrystallization is processed in a vacuum or an inert gas, so as to increase the grain size.
(c) After the fabrication of the target film, the target film and powder consisting of at least one of component elements forming the semiconductor compound for the fabrication of the target film are together contained in an airtight silica ampoule or the like to be heated at a high temperature. Usually, the saturated vapor pressure of the powder is higher than that of the target film, so that the powder penetrates into the inner portion of the target film in a gaseous condition, and then reacts with the deposits to be converted to the semiconductor compound.
(d) The inner parts of the target film is prevented from shortcircuiting by heating the target film in an atmosphere of oxygen or air so as to convert the deposits to oxides.
One of such the method is described in detail in "Crystallinity and Electronic Properties of Evaporated CdS Films", by J. Dresner et al., Journal of Applied Physics, Vol. 34, No. 8, August 1963, pp 2390-2395.
However, these conventional methods have in common a disadvantage that the fabrication process of the target is considerably complicated. Besides, in the above methods (a) and (b), the deposited film is easily separated due to the difference of expansion coefficient between the film surface and the glass substrate, so that the temperature cannot be increased and as a result, the increase of the grain size due to the recrystallization can hardly be expected.
On the other hand, the above methods (c) and (d) have disadvantages in that it is difficult to convert the deposits which locate in deep portions in the film from the surface thereof to a semiconductor compound completely, and that the non-uniformity or film separation occur frequently.
As mentioned above, the semiconductor compounds of groups IIb and VIb fabricated by the conventional high vacuum deposition method has various disadvantages even from the viewpoint of crystallography.
In addition, in the case where a photoconductive target of p-n junction type is fabricated by this method, there exist serious disadvantages from the viewpoint of energy level of the p-n junction, as follows.
Generally speaking, as a p-n junction which is preferable to a photoconductive target, the most preferable is a structure of "p.sup.+ -p-i-n-n.sup.+ " having a wide region of a p-i-n polarities and a narrow p.sup.+ region and a narrow n.sup.+ region which are overlayed on the both surfaces of the wide p-i-n region, respectively, as described later. Here, the p-i-n region has a p-type and an n-type semiconductor regions formed respectively on both sides of an intrinsic semiconductor, namely, an i-type region in which a Fermi level exists between a conduction band and a full band. The n.sup.+ region, in which a Fermi level exists in the vicinity of the conduction band, has a strong n-polarity. Similarly, the p.sup.+ region, in which a Fermi level exists in the vicinity of the full band, has a strong p-polarity.
In order to form the above-mentioned wide p-i-n region, it is required that a Fermi level exists as close as possible to the i-region in each of the p-region and the n-region. That is, it is required from a viewpoint of crystallography to reduce lattice defects as few as possible.
The photoconductive target is operated by applying a negative voltage to the p.sup.+ region and a positive voltage to the n.sup.+ region. Usually, the n.sup.+ region is in the form of a hole block, so that it has a structure of electron injection type. On the other hand, the p.sup.+ region is in the form of an electron block, so that it has a structure of hole injection type. Accordingly, even if the voltage applied across the p.sup.+ and n.sup.+ regions of the target is increased under this condition, electrons and holes are not injected from both sides, so that a dark current is extremely reduced. However, if the polarities of this structure are reversed due to contact potentials between the p.sup.+ and n.sup.+ regions and signal electrodes and contamination by a residual gas or the like (namely, the polarities are changed to n.sup.+ -p-i-n-p.sup.+), electrons and holes are injected into the film from the left side and the right side thereof, respectively, so that the dark current is increased extremely.
On the other hand, in the case where the regions of p.sup.+ and n.sup.+ are wide and the p-i-n region is narrow, the dark current itself is small. The p.sup.+ and n.sup.+ regions, however, act as electrode for applying an electrical field to the p-i-n region, and an electrical field for collecting photo-excited charges is not applied to the p-i-n region. Consequently, if the p.sup.+ and n.sup.+ regions are wide, the loss caused by an optical absorption is increased. That is to say, an amount of light arriving at the p-i-n region having a narrow photorensitive portion is reduced, and in addition the width of the p-i-n region is narrow or thin, so that the light arriving at the p-i-n region passes therethrough, and then disappears due to the absorption in the p.sup.+ and n.sup.+ regions. For these reasons, the loss of photo-electric conversion is increased extremely as a whole, so that the p.sup.+ -p-i-n-n.sup.+ structure can be the most preferable. These facts are well-known and are described in detail, for example, in "Photoconductive Properties of Lead-Oxide Layers", by L. Heijne, Philips Research Report, 1960, pp 1, 4, 5, 8, 9, 14, 15, 20, 21, 80-83, 106, 107, 148-151, or in "Concepts in Photoconductivity and Allied Problems", by Albert Rose, published by Interscience Publishers, 1963.
As apparent from the above-mentioned structures, in order to obtain an excellent photoconductive target formed by compounds of groups IIb and VIb, a new evaporation technique is required so that a semiconductor film having a desired thickness with desired semiconductor polarities is deposited on a desired position, i.e., at a desired depth from a surface of a film structure.
However, the evaporation technique as mentioned above has not been realized in a conventional high vacuum deposition method in case of p or n material of IIb and VIb compounds. According to the conventional method, a deposited film itself of IIb and VIb compounds show p.sup.+ or n.sup.+ polarity. If deposition material is p material, p.sup.+ film is obtained and in case of n material, n.sup.+ film is obtained. Consequently, a junction having the structure of n.sup.+ -p-i-n-p.sup.+ is apt to be fabricated. Were it possible to convert p.sup.+ or n.sup.+ film to p or n film by thermal treatment, it was still difficult to form a thin p.sup.+ or n.sup.+ region on an extremely narrow surface of those films. For instance, even if it is possible to convert a p region to p.sup.+ region by injecting an adequate impurity into the surface in accordance with an ion implantation method at each time of deposition or at each step of a deposition process, this conversion is not practical because of the complicated steps of the process.
Consequently, it is inevitably required that the polarity of a deposited film consisting of IIb and VIb compounds is controlled freely so that any desired polarity such as p.sup.30 , p, n, n.sup.+ or the like with a desired thickness during the deposition process.